Mass spectrometric and theoretical studies on protonated and potassium cationized biological molecules in the gas phase

Abstract

a-/b-amino acids are the constituent units of peptides and proteins, as well as important starting materials for the synthesis of pharmaceutically important intermediates and drugs. In particular, b-peptides are more resistant to enzymatic degradations than their a-analogues, and hold the key to a new generation of bacteria-resistance drugs. The recent surge in proteomics research also places high demands on peptide sequencing analysis by tandem mass spectrometry (MS/MS), which is based on the low-energy collision-induced dissociation (CID) mass spectra of protonated/alkali metal cationized peptides. However, the intrinsic properties of a-/b-amino acids/peptides, e.g. proton (H+) and potassium cation (K+) affinities, and the mass spectrometric dissociation mechanisms of protonated amino acids and peptides, are not fully understood. In the present study, the first set of reliable experimental K+ affinities of the twenty naturally occurring a-amino acids were determined to be (in kJ mol-1): Arg (163) > His (155) > Gln (154) > Tip (151) > Asn (149) > Lys (144) > Glu (140) > Tyr (140) > Phe (139) > Asp (137) > Thr (136) > Pro (135) > Met (135) > Ser (133) > Ile (129) > Leu (128) > Val (127) > Cys (124) > Ala (123) > Gly (119) using the mass spectrometric kinetic methods and estimated theoretically using the density functional theory (DFT) B3-LYP/6-311+G(3df,2p)//B3-LYP/6-31G(d) protocol. With the exception of lysine (Lys) and arginine (Arg), the experimental and theoretical potassium cation (K+) affinities are found to agree to within +-15 kJ mol-1, with mean absolute deviation (MAD) of 4.9 kJ mol-1 only. The most stable K+ binding to aliphatic amino acids (Gly, Ala, Val, Leu, Ile) involves a bidentate interaction in the CS1 form involving the carboxylic C=O and OH sites, whereas for amino acids with functionalized side chain (namely, Ser, Thr, Cys, Met, Phe, Tyr, Trp, Asp, Asn), a tridentate interaction in the CS2 form involving the backbone O=C, N-terminal NH2, and the O/N-heteroatom site of the functional group in the side chain is generally favored. The only exceptions are the K+-Pro/Glu/Gln/Lys/His/Arg complexes, which are found to be most stable in the zwitterionic ZW1 form. The H+/K+ affinities of five biologically important b-amino acids were determined to be (in kJ mol-1, at 298K/OK): b-Abu 942.0/134.7, b-Leu 955.6/138.3, b-Phe 948.0/137.6, b-Tyr 955.4/138.7 and b-Glu 951.5/141.5. The H+ affinities (PA) of four model b-dipeptides and one biologically important b-dipeptide were determined to be (in kJ mol-1 at 298K): Gly(b-Ala) 942.0, (b-Ala)Gly 971.3, Ala(b-Ala) 947.8, (b-Ala)Ala 970.3 and (b-Ala)His 1023.4. The experimental H+/K+ affinities for the b-amino acids and the H+ affinities for the b-dipeptides are found to be in very good agreement with values estimated by the DFT protocol, with MAD of 3.1/6.4 and 5.9 kJ mol-1, respectively. The most stable sites of binding and conformations of the H+ and K+ bound complexes were found by theoretical calculations using the DFT protocol. The most stable proton binding site of the b-amino acids and b-dipeptides is at the N-terminal amino nitrogen (NE2), while K+ binds to the two carboxylic oxygens [O=C and -OH] in the most stable charge-solvated CS1 mode. For carnosine ((b-Ala)His), the most stable proton binding site is at the pi-nitrogen of the imidazole ring of the histidine residue. The proton (H+) affinities (PA) of b-amino acids and b-dipeptides are generally larger than that of their a-analogues. This is attributed to a more stabilizing hydrogen bond or bonding network found in the b-amino acids and b-dipeptides. Unlike a-dipeptides, the PA enhancing effect is much more pronounced when b-Ala is located at the N-terminus than at the C-terminus of the b-dipeptides. The potential energy surfaces for the dissociation of protonated b-alanine (b-Ala), and two model b-dipeptides, (b-Ala)Gly and Gly(b-Ala), including the reaction intermediates, transition structures (energy barriers) and energetics of the reaction (Ho, H298, and G298) were found by theoretical calculations using the DFT protocol. For protonated b-Ala, the loss of H2O, CH2CO, and (CH2CO + H2O) pathways are energetically preferred. The loss of NH=CH2, NH3 and (NH3 + CO) pathways have higher critical energies, and are observed under more energetic CID conditions. The loss of these small stable neutrals are not found in the dissociation of protonated a-Ala, and is directly related to the presence of the 'extra' -CH2- unit in the main carbon chain of b-Ala, separating the N-terminal -NH2 group and C-terminal -COOH group. The formation of b-Ala specific fragment ions is dependent of the location, C-terminus versus N-terminus, of b-Ala in the b-dipeptide. For Gly(b-Ala), formation of the unique b'2 (oxazinone) ions is energetically favored over y1 ion formation. The b'2 (oxazinone) ion dissociates further by characteristic loss of NH=CH2 and NH3, in contrast to the loss of CO only observed for b2 (oxazolone) ions derived from a-dipeptides. For (b-Ala)Gly, formation of b1 (protonated b-lactam) ions and fragment ions due to the loss of NH3, which are indicative of the presence and location of b-Ala in the b-dipeptide, are competitive (having comparable energy barriers) with y1 ion formation. The theoretical findings are consistent with experimental observations, and could be used to rationalize the appearance threshold voltages and relatively abundances of different fragment ions in the CID-MS/MS spectra of protonated b-Ala, (b-Ala)Gly and Gly(b-Ala). The knowledge gained in the present study will be useful in sequence analysis of b-peptides by tandem mass spectrometry. The dissociation mechanisms of protonated a-peptides containing histidine is known to be complicated due to the presence and active participation of the basic imidazole ring in the side chain of histidine. In the present study, the potential energy surfaces for the dissociation of protonated model dipeptides, [GlyHis + H]+ and [HisGly + H]+, were established by DFT calculations. The formation of the b2(oxazolone-His) ion is energetically and entropically favoured, but b2(diketopiperazine-His) and b2(bicyclic) ions could also be formed at higher critical energies. As the b2(diketo-His) ion is much more stable, initially formed b2(oxazolone-His) ions could further isomerize to b2(diketopiperazine-His) ions if there is sufficient internal energy imparted to the b2 ions by collisional activation. The energy difference between the barrier of the formation of b2(oxazolone-His) and that of further isomerization to b2(diketopiperazine-His) ion are 201 and 115 kJ mol-1 for protonated GlyHis and HisGly, respectively. The difference in isomerization energy barriers is consistent with experimental observations in energy-resolved and time-resolved tandem MS/MS studies: isomerization is only partially achieved for b2 ions derived from protonated GlyHisGly, but complete isomerization is found for b2 ions derived from protonated HisGlyGly in the millisecond time frame of the ion trap mass analyzer. As the initial peptide sequence information is lost in the b2(diketopiperazine-His) ion (a protonated cyclo-peptide), the implication of b2 ion isomerization to peptide sequencing by tandem mass spectrometry is discussed.